baron
Solve a NLP/MINLP using the global MINLP solver BARON
Syntax
x = baron(fun,A,rl,ru)
x = baron(fun,A,rl,ru,lb,ub)
x = baron(fun,A,rl,ru,lb,ub,nlcon,cl,cu)
x = baron(fun,A,rl,ru,lb,ub,nlcon,cl,cu,xtype)
x = baron(fun,A,rl,ru,lb,ub,nlcon,cl,cu,xtype,x0)
x = baron(fun,A,rl,ru,lb,ub,nlcon,cl,cu,xtype,x0,opts)
baron(fun,A,rl,ru,lb,ub,nlcon,cl,cu,xtype,x0,opts,doSolve)
[x,fval,exitflag,info] = baron(fun,A,rl,ru,lb,ub,nlcon,cl,cu,xtype,x0,opts)
[x,fval,exitflag,info,allsol] = baron(fun,A,rl,ru,lb,ub,nlcon,cl,cu,xtype,x0,opts)
Description
x = baron(fun,A,rl,ru) optimizes a nonlinear function over a system of linear constraints using the global MINLP solver BARON. fun is a BARON compatible MATLAB function or anonymous function. A, rl and ru specify the linear constraints of the form rl ≤ Ax ≤ ru. To specify a one-sided constraint use Infinity on the empty side, and for equality constraints use rl = ru.
x = baron(fun,A,rl,ru,lb,ub) solves the linearly constrained problem subject to the decision variable bounds, lb and ub. Use -Infinity for an unbounded lower bound, and Infinity for an unbounded upper bound.
x = baron(fun,A,rl,ru,lb,ub,nlcon,cl,cu) optimizes a nonlinear function subject to the above-describd linear constraints, as well as the nonlinear constraints specified by nlcon, cl and cu of the form cl ≤ nlcon(x) ≤ cu. nlcon is a BARON compatible MATLAB function or anonymous function. To specify a one-sided constraint, use Infinity on the empty side, and for equality constraints use cl = cu.
x = baron(fun,...,cu,xtype) solves subject to the integer and binary constraints. xtype is a character array where 'C' are continuous, 'I' are integer and 'B' are binary variables.
x = baron(fun,...,xtype,x0) allows the user to specify an initial solution guess via x0. This is an optional field as BARON can generate its own starting point. To specify a partial starting point, fill unknown values in the x0 vector with NaNs.
x = baron(fun,...,x0,opts) allows the user to specify options specific to BARON via the structure opts. use baronset() to generate the options structure.
baron(fun,...,opts,doSolve) allows the user to just generate the output BARON model (mwrap.bar) and not solve the problem using BARON when doSolve = 0. The output file will be written to the current directory. The default is doSolve = 1 (solve the model).
[x,fval,exitflag,info] = baron(fun,...,opts) also returns the objective value at the solution, the solver exitflag and an information structure on the solver progress.
[x,fval,exitflag,info,allsol] = baron(fun,...,opts) also returns a structure containing all requested solutions and objective values when option numsol is greater than 1.
| Exitflag | Description |
| 1 | Optimal within tolerances |
| 2 | Infeasible |
| 3 | Unbounded |
| 4 | Intermediate feasible |
| 5 | Unknown |